End of 1990s stock market bubble period

The late-1990s stock-market bubble was not merely a period of inflated prices; it also altered the incentives that sustain trust. Executives at companies such as Enron and Tyco could reap enormous short-term gains while the future health of their businesses mattered less, investors grew careless about due diligence, and institutions meant to distinguish reliable companies from unreliable ones often failed without losing business. In the book’s wider argument, this period exposes a recurring problem: neither trust nor decentralized knowledge works well without systems that make information useful and misconduct costly.

Executive incentives and investor scrutiny

The bubble created what the book describes as a breeding ground for corruption by weakening the “shadow of the future”—the expectation that fair dealing and a company’s long-term success will matter later. Some CEOs knew their future cash flows could not justify their inflated stock prices. With abundant stock options and boards that paid little attention, the immediate rewards of self-interested conduct could outweigh concerns about the companies’ eventual prospects.

The contrast is especially stark in the example of Dennis Kozlowski, Tyco’s CEO. The book argues that it is difficult to see how he could have made $600 million honestly by remaining at the company, while dishonesty made such a sum remarkably easy. Investors had reason to recognize that the incentives for executives to honor promises or protect long-term business health had changed. Instead, intoxicated by bull-market gains, they neglected the due diligence that even trusting investors need to perform.

Watchdogs without consequences

The institutions that were supposed to limit corruption could instead help confer legitimacy on questionable companies. An underwriter’s involvement, an analyst’s buy recommendation, a stock exchange listing, or an auditor’s approval all signaled that a company had value or that its numbers could be trusted. Such signals depended on institutional credibility: in theory, firms that failed to identify problems would lose business, giving them a self-interested reason to do their jobs well.

That discipline did not operate as expected. Between 1997 and 2000, seven hundred companies were forced to restate earnings; in 1981, only three had been. The Nasdaq listed companies the book calls laughable, prominent firms such as Goldman Sachs underwrote them, and accountants approved their accounts. Yet the institutions responsible faced no corresponding marketplace penalty; they received more business, not less. Arthur Andersen had audited companies involved in major accounting failures, including Waste Management and Sunbeam, but investors did not treat the continued use of Andersen by companies such as WorldCom and Enron as a reason for skepticism. Investors stopped watching the watchmen, and the watchmen stopped watching too. The episode gives practical force to the principle “trust but verify.”

A wider problem of fragmented knowledge

The book places the bubble’s failures alongside cases in which a group’s information is divided among specialists or local actors. When the U.S. submarine Scorpion disappeared in May 1968 on its return toward Newport News after a North Atlantic tour, the Navy knew its last reported location but not what had happened or how far it had traveled. The initial search area was a circle twenty miles wide and thousands of feet deep. Rather than rely on a few leading experts, naval officer John Craven gathered people with varied knowledge, had them place separate wagers on possible scenarios, and used Bayes’s theorem to combine their judgments into an estimate of the submarine’s final location. The example illustrates how distributed estimates can be aggregated, rather than forced into a premature group consensus.

A different institutional response followed the intelligence failures preceding the attack on Pearl Harbor. In 1946, “Wild Bill” Donovan argued publicly for a centralized, independent agency to take charge of U.S. intelligence, citing both the prospect of conflict with the Soviet Union and new technologies such as rockets, atomic weapons, and bacteriological warfare. Congress created the CIA in 1947. Decades later, after the intelligence community failed to anticipate major terrorist attacks from 1993 through 2001, Congress again emphasized information sharing. Senator Richard Shelby criticized fragmented holdings and independent bureaucratic fiefdoms, arguing for an agency above those competing organizations. These contrasting cases sharpen the organizational question: centralization can bring separate information together, but it may also forfeit the independence and specialization that decentralization encourages.

Collective benefits and free-riding

Comparable coordination problems appear beyond financial markets. Television networks, stations, and advertisers could collectively benefit from more accurate local audience ratings, but no single participant had enough incentive to pay for an expensive alternative to paper diaries. Electronic people meters could track both programs watched and the viewers watching them; extending the system across the country’s 210 television markets might cost at least hundreds of millions of dollars, while roughly $25 billion in annual advertising spending relied on ratings gathered during sweeps. Each participant could hope another would fund an improved system and still share in its benefits. The result is the classic free-rider problem: a shared gain is available, but no individual actor finds it worthwhile to supply.

Taxation presents the same tension in a more fundamental form. Public goods, including roads and national defense, benefit people whether or not each individual pays toward them. That makes evading a contribution privately tempting, even though widespread evasion would erode the services everyone depends on. In 2001, the IRS audited only 0.5 percent of returns, so legal obligation alone could not explain high compliance. The book’s account emphasizes reciprocity: people are more willing to pay when they believe others are paying too and that evaders are likely to be caught and punished. The bubble’s collapse of scrutiny and the broader examples of fragmented information thus converge on a common point: collective outcomes depend not only on what individuals know or stand to gain, but also on whether institutions make cooperation credible and failure accountable.

Specialization, coordination, and organizational design

Decentralization has a genuine advantage: people closest to a problem often possess tacit knowledge tied to a particular job or place. But information discovered in one corner of an organization may never reach the rest of it. A group needs to preserve local expertise while finding ways to make it collectively useful—a balance that is as relevant to intelligence agencies and markets as it is to companies.

The book contrasts broken information channels with Toyota’s production system. Frontline workers are trained in a range of skills and organized into teams responsible for their production process; any worker can pull a cord to stop the line when something needs fixing. The cord is rarely pulled, but its availability gives workers authority over their immediate work and supports coordination without constant orders. Set beside the financial bubble, the example distinguishes effective decentralization from simple neglect: distributing responsibility can improve performance, but only when local knowledge can influence the larger system and those charged with oversight remain answerable for their decisions.